Scientific opinion on animal exposure to deoxynivalenol (DON) in feed materials and compound feed in the Republic of Croatia

The Croatian Agency for Agriculture and Food (HAPIH), through its Centre for Food Safety (CSH), has a statutory role in providing scientific and technical support to the competent authorities and in preparing scientific opinions on risks in the food and feed. In this context, the risk assessment of mycotoxins represents an important area of activity, with the mycotoxin deoxynivalenol (DON) standing out as one of the most frequently occurring contaminants in animal feed. DON belongs to the group of trichothecene mycotoxins produced by fungi of the Fusarium genus and is one of the most commonly detected mycotoxins in cereals and products of plant origin. Exposure of animals to DON may reduce feed intake and adversely affect growth, health, and production performance, with pigs and poultry being among the most sensitive species. Owing to its frequent occurrence in feed materials and compound feed, as well as its potential adverse effects on animal health, the assessment of exposure to and risk from DON is of particular importance in the field of feed safety. This scientific opinion on DON was prepared using data from official feed controls in the Republic of Croatia covering the period 2020‐2024, as well as HAPIH data collected under Measure 14 Animal Welfare and Intervention 70.06. The exposure assessment was carried out for fattening cattle and heifers, breeding heifers, dairy cows, dairy goats, fattening lambs, fattening pigs, broiler chickens, laying hens, and fattening turkeys. Analysis of the available data revealed considerable variability in DON concentrations among samples, including the occurrence of individual highly contaminated samples, indicating the heterogeneous nature of contamination with this mycotoxin. At the same time, certain limitations were identified, particularly those associated with non‐harmonised feed classification and the presence of left‐censored (LC) data. The exposure assessment was performed according to the methodology of the European Food Safety Authority (EFSA), using data on DON concentrations in feed, feed intake, and animal body weight. To account for uncertainty associated with LC data, DON concentration data were processed using the lower‐bound (LB), middle‐bound (MB), and upper‐bound (UB) approaches. Results from both datasets showed that the mean DON concentrations in all analysed feed materials and compound feed were below the applicable guidance values. Maize was identified as the most important source of DON exposure for the majority of the animal species and categories considered, while the HAPIH dataset additionally highlighted total mixed ration (TMR), whole‐crop maize silage, and certain compound feeds as important contributors to exposure. For most animal species and categories assessed, the estimated exposure levels were below the toxicological reference values established by EFSA. Ruminants, including fattening cattle and heifers, breeding heifers, dairy cows, dairy goats, and fattening lambs, were considered to be at low risk, as their estimated exposure levels remained below the reference values. The highest estimated exposure was observed in poultry and fattening pigs. In the official control dataset, the estimated exposure levels for broiler chickens and fattening turkeys exceeded the EFSA reference values. For fattening pigs, the estimated exposure levels did not exceed the reference value but approached it, indicating the need for continued monitoring. Exposure estimates based on the HAPIH dataset showed lower levels of exposure, and no exceedance of the reference values was identified for any of the animal species and categories assessed. The observed differences between the two datasets are most likely attributable to differences in data sources, sampling design, and the characteristics of the analysed samples. The uncertainties associated with the exposure assessment relate to the observed variability in DON concentrations, the use of average feed intake and body weight values, limitations in feed ration composition data, and the representativeness of sampling. Despite these limitations, the exposure estimates are considered sufficiently reliable to support scientific conclusions. The results of this scientific opinion indicate that DON currently does not pose a significant risk to most of the animal species and categories considered in the Republic of Croatia. However, the identified high‐exposure scenarios suggest a potential risk for poultry and, to a lesser extent, for fattening pigs. Consequently, systematic monitoring of DON occurrence in feed should be continued, and risk assessments should be regularly updated as new data become available. Particular attention should be given to feed materials and compound feeds identified as the major sources of exposure, particularly maize. Among other measures, it is recommended to standardise feed categorisation and harmonise nomenclature across databases, reduce the proportion of left‐censored data through the application of more sensitive analytical methods, and incorporate information on actual feeding practices and the geographical origin of samples in order to improve the accuracy of exposure assessments. Although current knowledge indicates that the carry‐over of DON from feed into food of animal origin is generally very low, continuous monitoring of DON levels in feed materials, compound feed, and food of animal origin remains important for the timely identification of potential changes and for maintaining food safety.